More about HKUST
Using mobile crowdsourcing data to retrieve the location-aware information
PhD Thesis Proposal Defence Title: "Using mobile crowdsourcing data to retrieve the location-aware information" by Mr. Xiaonan GUO ABSTRACT: Location-based services (LBS) can benefit from location information and indoor maps Recent research focuses on how to use sensor reading from smart phone to automatically reconstruct the walking pathway and floor plan. However, existing approaches have two limitations. On the one hand, the floor the room boundaries detection algorithm used in previous work may regard counter, shelf or obstacles as wall. On the other hand, having outline information may not be very useful to the commercial expense activities, such as advertising, shopping and dining. In this proposal, we aim to design a system that makes use of dynamic crowdsourcing data from smart phone to accurately construct the floor plan of shopping malls and label shop with types and brand name. Our empirical experiments show that people moving pattern and acoustic information are varied from different shops. Moreover, due to the pervasive WiFi, nearly all the shops have APs to provide wireless connections either for customers or their own employee and they are more likely to use the brand name as SSID. By investigating these observations, we design a novel gradient based algorithm in shops boundaries detection. Then we leverage people moving pattern and acoustic information obtained from smart phone to classify shops into different types. Finally, we generate a WiFi heat map from crowdsourcing data and matching APs in the floor plan to pinpoint shop location and use SSID to indicate brand name. We only leverage mobile phone sensor reading without human intrusion. Date: Wednesday, 8 May 2013 Time: 1:00pm - 3:00pm Venue: Room 3405 lifts 17/18 Committee Members: Prof. Lionel Ni (Supervisor) Dr. Huamin Qu (Chairperson) Dr. Lei Chen Dr. Qiong Luo **** ALL are Welcome ****